Aluminum from clay: two processes developed by the bureau of

World War II. Shortly after the ... submarine warfare in the Caribbean seriously inter- fered with ..... the mixture is filtered anti an extract conta...
0 downloads 0 Views 4MB Size
MARCH, 1948

ALUMINUM FROM CLAY Two Processes Developed by the Bureau of Standards

W~mnmresearch a t the National Bureau of Standards has gone far in solving a problem that has been the object of numerous investigations by chemists over a period of many years-that of producing aluminum from clay. The development of two processes a t the Bureau makes it possible that this abundant raw material may sometime replace bauxite as a source of the metal. Actually, the manufacture of aluminum from clay will in the near future be imperative for the protection of the United States from dependence upon foreign sources of aluminum ore. According to Interior Department estimates, 70 per cent of the supply of domestic bauxite has already been consumed. Under normal consumption, it is calculated, there is sufficient bauxite in this country to last only nine years. Because of the limited domestic supplies of low-silica bauxite, about 70 per cent of the needs of the United States aluminum industry was being importcd long before our entry into World War 11. Shortly aftpr the United States delcared war in 1941, submarine warfare in the Caribbean seriously interfercd with bauxite shipments, and it became important to utilize dompstic sources of aluminym, of which clay is the most abundant. Bauxite (~mpure hydrated alumina) tmd clay (mainly hydrated aluminum silicate) both may scrvc as sources of purified alumina (A1208), from which the aluminum mctal is obtained by electrolytic reduction. Accordingly, in 1941 investigations were brgun at thr National Bureau of Standards seeking economically feasible methods for the production of alumina from clays and low-grade bauxites. This work

resulted in the development of two processes for the recovery of alumina from clay-one a hydrochloric acid extraction process, the other an alkaline extraction with a solution consisting of a mixture of s d u m hydroxide and sodium chloride. In the last three-quartem of a century, at least sixty processes had been proposed for deriving alumina from common clays and other aluminous materials besides bauxite. Most of these processes were found impractical, howkver, because of the high cost and low purity of the product. When alumina is reduced electrolytically to metallic aluminum, the impuritied, largely silica and ferric oxide, also are reduced and pass into the metal. It is thc requirement that the iron and silicon in the aluminum be kept low that demands the use of highly purified alumina. The various methods proposed for the recovery of alumina from clays and high silica bauxites can be divided into two groups-alkaline processes and those requiring the use of acids. In both groups the problems of obtaining a high recovery of alumina and of cffectively separating silica and ferric oxide from the alumina must be solved. In acid processes there has been the additional problem of corrosion of equipment. ACID EXTRACTION

The acid extraction developed a t the National Bureau of Standards is the only successful hydrochloric acid process that has been carried to completion. The basic reactions involved had hrcn usrd and studied in the laboratories of the Bureau in the analysis of bauiite, high-alumina refractories, aluminum metal, and mis-

JOURNAL OF CHEMICAL EDUCATION

cellaneous ceramics, ahd in the redetermination of the atomic weight of aluminum. The problem was to adapt the laboratory methods to a commercial process. Previous attempts a t development of a hydrochloric acid process had failed because of the corrosion of metallic equipment by acid fumes. This problem, as well as the abrasive effect of precipitated aluminum chloride crystals, was overcome in the pilot plant through the use of modern plastics, glass, and refractory materials in place of metals. Glass piping had the additional advantage of making certain stages of the process visible to the observer. Preliminary experimental work had centered around two major problems-the decomposition of clay with hydrochloric acid and the efficient separation of aluminum from iron and other soluble constituents of the clay: Alumina can he dissolved out of all clays by treatment with hydrochloric acid in closed containers at 300°C., but the high pressures developed rendered this method of decomposition unsuitable for plant installations with the equipment available,. . However, it was found that roasting the clay at the relatively moderate temperature of 700°C. furnished enough energy to break the chemical union between aluminum and silicon in the aluminum silicate molecule. It was then possible to dissolve out the alumina with dilute hydrochloric acid at about 100°C. under atmospheric pressure, leaving the silicon in the form of silica, insoluble in the acid. The problem of the separation of aluminum from iron and other acid-soluble constituents of clay re: mained. Remov?l of the aluminum by precipitation with hydrochloric acid gas as the hydrated chloride was found most practical, since almost all of the impurities are left in solution. The hydrochloric acid process consists in (1) roasting the clay; (2) digesting the roasted product in dilute hydrochloric acid; (3) filtering to separate the'insoluble siliceous matter from the solution containing the aluminum and soluble impurities such as iron and alkali salts; (4) concentrating the solution; (5) precipitating the alnminum as the hydrated chloride from the concentrated solution by adding hydrochloric acid gas; (6) removing the crystals of hydrated aluminum chloride; (7) washing the crystals to remove adhering impurities; (8) calcining the hydrated chloride to obtain alumina; and (9) recovcring hydrochloric acid from the waste products at the end of the process. While the process is applicable to nearly all clays, kaolin was used in the pilot plant because of its abundance andaccessibility. The raw material first cntcrs a triple-hearth Wedge furnace, where it is roasted at ahout 700' C. to incrcasc the solubility of its alumina contcnt iu acid. I t then passcs into a digester, in which it is stirrcd to mix with hot dilutc hydrochloric acid until solubla material is dissolved. For commercial operation, digestion for one hour a t %boot 100°C. is sufficient.. T l ~ slurly r in thv digrst,er is t,hm coolcd t o about 60% by turning off the steam and circulating cooling water through the heaters.

From the digester the cooled slurry is pumped into a filter press where the insoluble matter, consisting largely of silica sludge, is removed. After all of the slurry is pumped into the press, the filter cake is blown dry with compressed air and then washed with water. The acid filtrate and washings pass into a still, in which the solution is evaporated to a relatively small volume to minimize the quantity of hydrochloric acid gas required in the next step to precipitate the aluminum as the hydrated chloride (AlCl3.6Hz0). The water distillate is discarded, while the acid distillate is retained for fut,ure decomposition of clay in the digester. The concentrated solution is then cooled and enters a precipitator, where intimate contact with hydrochloric acid gas results in the precipitation of the hydrated aluminum chloride. This is accomplished by pumping the solution to the top of an unpacked tower, where it is allowed to fall against a stream of hydrochloric acid gas entering below. Because of the heat envolved in the absorption of the hydrochloric acid gas, the solution is circulated through a cooler during the process to increase the absorption rate. From the precipitator, the aluminum chloride, suspended in a solution saturated with hydrochloric acid,. goes to a rubber-coated centrifuge with perforated basket, where the chloride crystals are removed from the solution by filtration through an asbestos cloth. Impurities adhering to the aluminum chloride are removed by washing with a hydrochloric acid spray. Conversion of the hydrated aluminum chloride to the, oxide is accomplished by calcination in a muffle-fired furnace with constant agitation of the aluminum chloride by the rabble arms as it passes downward over the heated hearths. This agitation makes the conversion of chloride to oxide a continuous process and effects the conversion at a lower temperature than would otherwise be necessary. A rubber-coated exhaust fan pulls the hydrochloric acid gas and water vapor from the top of the furnace. The acid gas is drawn through a cooler, where a certain amount of the aqueous acid condenses and is drained into a still for later generation of hydrochloric acid gas used in the process. The major portion of the acid gas, after passing t,hrough the cooler, is absorbed in cool hydrochloric acid of constant boiling strengt,h in an absorption tower by the same technique as that used in the precipitation of aluminum chloride. In this way the constant boiling acid is built up to acid of sprcific gravity 1.18 to 1.20, which is then pumped into a still and heated to produce hydrochloric acid gas for prccipitation of aluminum chloride. The alumina emerging from the bottom of the furnace is removed by a vibrating conveyor so arranged and regulated that the lower end of a 6-inch glass outlet pipe is always filled with oxidr. This prcvmts air from entering t,hr furnace and diluting t,hchydrochloric acid gas issuing from the top. The alumina obtained by this method has an average purit,y of about 90.8 per cmt,. Thr only significant, impurities are 0.1 per cent of chlorine, 0.04 per cent of iron oxide, and 0.06 per cent of silica. If a purer prod-

uct is dcsircd, the chloride content can be further reduced by longer heating at higher temperatures, and the content of iron oxide can readily be lowcred by additional washing of the chloride in the centrifuge basket. In addition to its use in production of aluminum metal, the alumina produced by this process is superior to any on the market for polishing metallographic specimens. Exceptionally good results have also been obtained with this alumina in the preparation of enamels for use a t high temperatures, probably due to the extreme fineness of the product. The price of producing aluminum by the hydrochloric acid process is at present about twice that from imported bauxite. However, studies now in progress a t the Bureau may result in the reduction of the original cost figures. The two major cost items are fuel for heating and the replacement of hydrochloric acid, of which the latter is the more important. Greater efficiency in the re-use of hydrochloric acid offers possibilities of further economy, as does the better use of power through more efficient heat exchangers and the utilization of waste heat from furnace gases. There are also possibilities in the recovery of usable by-products, such as iron. from waste solutions. ALKALINE EXTRACTION

Because a fund of information concerning the behavior of alumina in alkaline solutions had been built up as a result of studies not directly concerned with the recovery of alumina, an investigation of alkaline processes was also undertaken a t the National Bureau of Standards. The key results of this work, representing a suggested commercial process for the extraction of alumina from clay and high-silica bauxites, were available by early 1943. However, before the process could be carried out on a larger scale, it appeared that the submarine menace in the Caribbean might be controlled, and in October, 1943, domestic production of bauxite was slashed nearly in half. Thereafter, during World War I1 dependence for the production of alumina was placed in bauxite convoyed from South America. As a result, the alkaline process was never taken to the pilot plant stage. One of the fundamental principles of many of the alkaline processes is the conversion of the alumina in the ore to sodium aluminate, which is readily soluble in water. The iron is easily removed, since it is very insoluble in alkali, but considerable silica remains in solution. The problem is to rid the solution of the bulk of the soluble silica, so that the silica content of the precipitated alumina may be below an acceptable limit. The investigation, therefore, was first concentrated on this problem. I n the extraction of alumina from ore as sodium aluminate, there are three principal methods: (1) direct extraction of the untreated ore with sodium hydroxide solution as in the Bayer process; (2) treatment of the ore to give a product containing solid sodium aluminate, which is extracted with water; and (3) treatment of the ore t o give a product containing

.

calcium aluminate, which can be decomposed with sodium carbonate solution. Each of these mcthods rcsults in a sodium aluminat,e solution containing solublc silica t,o the extent of one or two pcr ccnt of the alumina present. I t is necessary, however, to reduce the percentage of silica below these values. As a point of departure in the study of these alkaline methods, numerous extractions were made with samples of domestic high-silica bauxite, using sodium hydroxide solution a t boiling temperature. As expected, solutions of sodium aluminate containing moderate amounts of silica were obtained. Analyses of the insoluble extraction residues showed that soda, alumina, and silica were present in simple, constant molar proportions, indicating that the residues might consist principally of a single compound. Accordingly, attempts were made t o synthesize from pure starting materials a sodium aluminum silicate compound similar in composition to the extraction residues. X-ray diffraction patterns of the resulting products strongly resembled those of the naturally occurring mineral sodalite, 3NaO.3Al2Oa.6SiO2.2NaCI. As a result of these observations, further extractions of the bauxites were made usinn sodium chloride in addition to the sodium hydroxide,and the silica content of the extracted solutions was immediately reduced to one-tenth or less of the former value. The reduction in residual silica was caused by the formation during the extraction of a compound of the sodalite type having a significantly lower solubility than the sodium aluminum silicate formed in the absence of sodium chloride. This was an important discovery, for a t once it presented a method of reducing the silica in alkaline extraction solutions to a point at which alumina suitable for electrolytic reduction could be precipitated with carbon dioxide. Numerous extractions were performed to determine optimum conditions, and it was found that only about 75 per cent of the alumina in the impure domestic bauxite could be dissolved out with sodium hydroxidesodium chloride solutions, because the remainder of the alumina formed an insoluble hydrated sodium aluminum silicate. At the same time, soda was lost in the formation of this complex. Attention was therefore directed to recovering the precipitated alumina and soda. The extraction residues were sintered with lime, and the resulting solid sodium aluminate was extracted with water. Sodium aluminate solutions were thus obtained containingmost of the soda lost in the 6rst extraction and sufficient alumina to make the over-all extraction 90 per cent or better. Silica was present only t o the extent of one or two per cent of the alumina and could be further reduced t o a very low value by boiling the solution in contact with a seed charge of sodalite in the presence of sodium chloride. Most important, it was discovered that the presence of sodium chloride had no adverse effect on the extraction of the alumina and soda from the residues. This meant that the spent solution from the precipitation of alumina with carbon

162

dioxide could be recirculated and used for extracting thc sintered residue. The results of these investigations on high-silica bauxites, in which the silica was reduced to a low value by the application of the sodalite principle, led to experiments seeking a method for the more complete separation of silica from alumina in solutions obtained in the lime-sinter process. Ground siliceous aluminum ore was mixed with finely divided calcium carbonate in the proper proportions to give a material equivalent t o a mixture of dicalcium silicate with pentacalcium trialuminate and mono- or tricalcium aluminate. This mixture was then heated at temperatures approaching 1400°C. for the time necessary for formation of these compounds. The resulting sinter was cooled slowly to about 1200°C. to permit crystallization of any liquid present at the higher temperature. Thereafter the material was cooled to room temperature without particular attention to cooling rates. The dicalcium silicate played an important dual role in this procedure. First, it held the silica in an unreactive form insoluble in the alkaline extracting solution. Second, it expanded on cooling with consequent reduction of the sintered mass t o a powder--a process known as "dusting," which rendered the aluminous materials accessible to the action of an alkaline solution without grinding. I t was found that sodium chloride in the extracting solution did not seriously interfere with the extraction of the alumina with sodium carbonate solution. This detail permitted the further use, as in the treatment of lime-soda sinters from bauxite extraction residues, of the spent solutions from which the silica and alumina had been removed by precipitation with sodalite and carbon dioxide, respectively. These experiments-were the basis for the establishment of a method of recovering alumina from clay by a lime-sinter process involving extraction with a sodium carbonate-sodium chloride solution. The lime-sinter process, by no means new in general outline, was improved by utilization of the sodalite method of desilicating alumina solutions containing silica. A means was thus afforded for the production from clay of hydrated alumina, which after calcining was entirely suitable for electrolytic reduction to the metal. From the available information, flow sheets for suggested commercial processes were prepared. I n the outlined process for the alkaline extraction of alumina from clay, a mixture of clay and limestone is prepared by either wet or dry process grinding and mixing. Commercial equipment is available that appears to be particularly well adapt,ed for burning the raw mix and annealing the resulting clinker. This apparatus con-

JOURNAL OF CHEMICAL EDUCATION

sists of an ordinary rotary kiln in series with a rotary cooler, the lat.ter carrying a burner installed a t its feed end. Thus the clinker discharged from the primary kiln can be annealed in the cooler at a rate determined by its speed of rotation and by adjustment of the burner. Complete dusting of the annealed clinker reduces the need of further grinding. A solution of sodium carbonate and sodium chloride is added to the powdered mixture, and the resultant slurry is stirred a t high speed for about one-half hour a t 60 to 80°C. On completion of the extraction period, the mixture is filtered and an extract .containing 70 to 80 grams of alumina and 1to 2 grams of silica per liter is obtained. By subjecting the residue to couutercurrent washing with salt solution, it is expected that the concentration of alumina in the wash water can he built up to approach that in the extract. The residue, consisting principally of lime and silica, may then be washed with water and stock-piled as a raw material for the production of portland cement. The filtered alumina solution plus concentrated washings is next desilicated by boiling for one hour with a seed charge of synthetic sodalite, which induces the precipitation of crystalline sodalite. After filtration, the sodalite in excess of that required for seed for the next batch is returned to the first stage of the process. In this way the alumina and soda in the excess sodalite are recovered. Hydrated alumina is precipitated by passing carbon dioxide, obtained from scrubbed kiln gases, into the desilicated solution. I n order t o avoid contamination by silica, the precipitation of alumina is not carried to completion. However, in cyclical operation of the process complete recovery of the alumina from solution would be obtained. The alumina is filtered and washed with water low in silica to avoid contamination, and the residual solution and washings are used to scrub the kiln-flue gases. This procedure causes not only removal of dust from the gases but also evaporation of water from the combined solution and washings. Sufficient make-up quantities of sodium carbonate and sodium chloride are added to the evaporated and carbonated solution, so that it may be used for the extraction of a fresh batch of sinter. The hydrated alumina is dried and calcined for electrolvtic reduction t o aluminum. About 95 per cent of the alumina in the clay is recovered by this method, while losses of soda are small. An advantage of the process is that all steps can be conducted at atmospheric pressure, whereas many other alkaline methods for the extraction of alumina require autoclave desilication.